Installation Guide

There are various methods to run the application in production. We only support the Docker/docker-compose setup. If you plan to run the application directly on a host-system you should know what you are doing. If you are looking for a development setup please take a look at the Development section.


For production usage you should run the Code FREAK backend application and the dedicated Docker daemon it connects to on different hosts.

Backend application:

Because the backend application is provided as Docker image you need a running installation of Docker on your server. The application itself is not demanding many resources. A single CPU core and about 1GiB of RAM should be sufficient.

Dedicated Docker host:

The dedicated Docker host is used to run the web IDE instances and the evaluation. Because the containers will run arbitrary untrusted code from students you should take additional security measures to secure this machine. Especially the VSCode web IDE will require lots of memory (depending on the programming language). Some tests have shown that Java projects can consume more than 3GB of RAM per user. Simple C projects will only require about 500MB RAM. Depending on the programming languages you should plan the required resources accordingly.

We are working on an alternative IDE that reduces the resource usage per user.

Postgres Database

Code FREAK uses Postgres to store its data and files. You can either use a dedicated/managed database host or start a database container next to the backend application. Check out the configuration docs on how to configure the database connection.

Preparing the Docker daemon

The dedicated Docker instance will run on a different machine. Docker exposes an API that Code FREAK speaks to. By default it will search for a UNIX Domain Socket located at /var/run/docker.sock. If you run the CF backend and Docker host on two different machines you will have to enable the HTTP API of Docker. The communication than runs over TCP and should be secured by TLS. There is an extensive documentation on how to do this in the official Docker docs. Please follow this article carefully. As the result you should have a ca.pem, cert.pem and key.pem file that can be used on your backend application host. You need these files for the following steps.

Docker Compose example configuration

You could spin up Code FREAK with docker run but for a reproducible setup you should utilize a container orchestration system like Docker Compose. The following is a example production configuration for Docker Compose that runs the backend application and a Postgres database container:

version: "3.7"
    image: "cfreak/codefreak"
    restart: "on-failure:10"
      # you might want to use a reverse proxy that passes the traffic to Code FREAK
      - "80:8080"
      - default
      - private
      - "./application.yml:/app/resources/config/application.yml" (1)
      - "./certificates:/var/codefreak/certs" (2)
    image: "postgres:11"
    restart: "on-failure:3"
      - private
      - "postgres_data:/var/lib/postgresql/data"

  private: {}

  postgres_data: {}
1 We mount an application.yml file into the container. This allows us to configure Code FREAK. Below is a minimal configuration file that uses the database container and the dedicated Docker host.
2 Place the ca.pem, cert.pem and key.pem files from the previous step in a directory certificates next to your docker-compose.yml.
# Configure the Postgres database
    url: "jdbc:postgresql://postgres:5432/postgres"
    username: postgres
    # We enabled "trust" authentication for our Postgres DB. This requires no password
    #password: <none>
    driver-class-name: org.postgresql.Driver
    database-platform: org.hibernate.dialect.PostgreSQLDialect
      ddl-auto: validate
        globally_quoted_identifiers: true

    host: "https://[your-docker-host-url]:2376"
    cert-path: /var/codefreak/certs


For further information and available settings please check out the Configuration section of the docs.